No relevant resource is found in the selected language.

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy>

Reminder

To have a better experience, please upgrade your IE browser.

upgrade

IGP and BGP Route Flapping Occurred Due to Incorrect Setting of IGP Route Preference

Publication Date:  2013-09-27 Views:  26 Downloads:  0
Issue Description

Networking:

Backbone A------------------Backbone B

  |                                             |

  |                                             |

NE5KE(A)-------------------NE5KE(B)

 

Metro network equipment NE5KE and provincial backbone network equipment interconnected through EBGP. The NE5KEs in the metro network advertised routes to each other using OSPF, an IGP type, and advertised routes to the backbone networks using BGP.

Symptom: After services were migrated to NE5KE-A, some users complained that their E-Line services were faulty and Internet access speed was slow.
Handling Process

Possible causes include:

1. The forwarding performance of NE5KEs degraded.

2. Certain links did not work properly.

3. Routing problems existed.

 

Huawei completed the following steps to diagnose the problem:

1. Checked FIB table entries at the software level and hardware level. No exception existed

2. Queried port data. Error packets did not increase. Ping tests indicated that links were available.

3. Checked whether NE5KE-A learned service routes within the metro network.
NE5KE-A learned the routes of the service network segment (XXX.YY.24.0/23) connected to SRs. The routes were OSPF ASE routes and egress port g2/0/8 was toward the SRs.

<A>dis ip ro XXX.YY.24.0

Routing Table : Public

Summary Count : 1

Destination/Mask    Proto  Pre  Cost       NextHop         Interface

    XXX.YY.24.0/23  O_ASE  70   21         WWW.ZZZ.251.10  GigabitEthernet2/0/8

 

4. Checked the service network segment of the metro network advertised from NE5KE-A to the backbone network.
Service network segment XXX.YY.24.0/22 was advertised over a blackhole route with a 22-bit mask through BGP.

ip route-static XXX.YY.24.0 255.255.252.0 NULL0 preference 210----The preference value of the blackhole route was 210.

ipv4-family unicast

  undo synchronization

  preference 20 80 80

  network XXX.YY.24.0 255.255.252.0-----The network segment advertised by the BGP network had a 22-bit mask.

 

5. Checked how the NE5KE-A (NE A) and NE5KE-B (NE B) imported static routes.
They imported all static routes through OSPF.

ospf 163 router-id y.y.y.y

 default-route-advertise cost 1 type 1

 import-route direct cost 0 type 1

 import-route static type 1-----No filtering policy was set for importing static routes.

 preference ase 70----The preference value of OSPF ASE routes was 70.

 

6. Checked the service route XXX.YY.24.0, which had a 22-bit mask, on NE A and NE B. Route flapping occurred, with the blackhole route, OSPF ASE route, and IBGP route alternately selected.

Preference value of varied routes:
Static routes: 210
OSPF ASE routes: 70
IBGP routes: 80
The following describes how the route flapping occurred.
I. The same static blackhole route (preference: 210) was selected on NE A and NE B.
II. NE A advertised the blackhole route to NE B through OSPF ASE.
NE B compared the preference value of the imported route (preference: 70) with that of the native static blackhole route, and selected the imported OSPF ASE route.
III. NE B advertised the blackhole route learned from the backbone network, which had learned the route from NE A through BGP (BGP advertisement did not change the preference value of blackhole route, which remained 210), to NE A through IBGP.
NE A compared the preference value of the imported route (preference: 80) with that of the native static blackhole route, and selected the imported IBGP route.
IV. After the native static blackhole route on NE A was ineffective, the dynamic routes derived from the static blackhole route lost effect:
----On the NE B side, the OSPF ASE route (preference: 70) imported from NE A and the IBGP route (preference: 80) imported from the backbone network lost effect.
----On the NE A side, the IBGP route (preference: 80) imported from NE B lost effect.
V. The same static blackhole route (preference: 210) was selected on NE A and NE B again.
This cycle repeated itself, meaning that BGP routes kept flapping.
Root Cause
IGP and BGP routes flapped.
Solution

1. Set the preference value to 50 for static routes and to 150 for OSPF ASE routes.

2. It is recommended to modify the OSPF route import policy, so the static blackhole route is not imported.
Suggestions
Similar problems had frequently occurred, because a wide range of routing protocols and policies is involved in the egress direction of metro networks. When configuring routing policies for metro networks, pay special attention to the priorities of varied protocols and policies.

END